Basic Research—Biology

ADAM28 Manipulates Proliferation, Differentiation, andApoptosis of Human Dental Pulp Stem CellsZheng Zhao, PhD,*† Hongchen Liu, PhD,* and Dongsheng Wang, PhD*

Abstract

Introduction: The purpose of this study was to investi-gate the influence of a disintegrin andmetalloproteinase28 (ADAM28) on the proliferation, differentiation, andapoptosis of human dental pulp stem cells (HDPSCs)and possible mechanism. Methods: After ADAM28 eu-karyotic plasmid and antisense oligodeoxynucleotides(AS-ODNs) were constructed and respectivelytransfected into HDPSCs by Lipofectamine 2000, theADAM28 expression levels among diverse groupswere estimated by reverse transcription polymerasechain reaction (RT-PCR) and western blotting. Metha-benzthiazuron (MTT) and cell cycle assays were usedto test the HDPSCs proliferation activity. Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide andalkaline phosphatase analysis were performed respec-tively to measure apoptosis and the cytodifferentiationlevel. Immunocytochemistry and western blotting wereperformed to determine the effects of ADAM28 eukary-otic plasmid on HDPSCs expressing dentin sialophos-phoprotein (DSPP), dentin matrix protein 1, and bonesialoprotein. Results: ADAM28 could be correctly tran-scribed, translated, and expressed in HDPSCs. TheADAM28 AS-ODN group displayed the highest opticaldensity value, whereas the eukaryotic plasmid groupshowed the lowest, which suggested that ADAM28had a negative regulatory effect on the proliferation ofHDPSCs. ADAM28 eukaryotic plasmid could significantlyinhibit the HDPSC proliferation, promote specific differ-entiation of HDPSCs, induce apoptosis, and enhance theDSPP expression, whereas ADAM28 AS-ODN producedthe opposite effects. Conclusions: Our results provedthat ADAM28 might actively participate in manipulatingthe proliferation, differentiation, and apoptosis ofHDPSCs. (J Endod 2011;37:332–339)

Key WordsADAM28, apoptosis, differentiation, human dental pulpstem cells, proliferation

From the *Institute of Stomatology, General Hospital of ChineseChinese People’s Liberation Army, Qingdao, Shandong, China.

Supported by grants from the National Nature Science FoundatAddress requests for reprints to Dr Hongchen Liu, Institute of Stom

100853. E-mail address: zhaozheng6912@sohu.com0099-2399/$ - see front matter

Crown Copyright ª 2011 Published by Elsevier Inc. on behalf odoi:10.1016/j.joen.2010.11.026

332 Zhao et al.

It has been shown that human dental pulp contains a rapidly proliferative subpopula-tion of precursor cells termed dental pulp stem cells (DPSCs) that show self-renewaland multilineage differentiation into odontoblast-like cells secreting a mineralizedmatrix having the mineral and molecular characteristics of dentine and also secretemultiple proangiogenic and antiapoptotic factors (1). Human dental pulp stem cells(HDPSCs) reside predominantly within the perivascular niche of dental pulp and arethought to originate from migrating neural crest cells during development (2). More-over, HDPSCs are considered to give rise to dentin-pulp complex-like structures ora woven bone-like structure in vivo and to harbor great potential for tissue-engineering purposes (1, 3).

A disintegrin and metalloproteinases (ADAMs) are a new gene family of proteinswith a sequence similar to the reprolysin family of snake venomases that share the met-alloproteinase domain with matrix metalloproteinases (MMPs) (4), and about onethird of the family members have the catalytic site consensus sequence in their metal-loproteinase domains (5). A disintegrin and metalloproteinase 28 (ADAM28), asa newly discovered gene, is composed of several domains including propeptide, metal-loproteinase, disintegrin, cysteine-rich, epidermal growth factor (EGF)-like, transmem-brane, and cytoplasmic tail domains (5).

Previous studies showed that ADAM28 is overexpressed selectively in the non–small-cell lung carcinomas and is considered to promote breast carcinoma cell prolif-eration through enhanced bioavailability of IGF-1, thus suggesting the possibility thatADAM28 is involved in cancer cell proliferation, metastasis, and progression (6).

Several years ago, the ADAM28 gene was screened from patients with congenitalhypoplasia of tooth root (CHTR) by Zhao et al (7). Research displayed that CHTR isa disease characterized by a physiological development disorder of tooth root resultingfrom ectodermal dysplasia, and the patients usually present with tooth mobility, atoniamasticatoria, and early slough (8). Nowadays, no effective therapy has been found in theworld, and its potential mechanisms are not clear. Thus, ADAM28 has been regarded asone of the possible virulence genes for CHTR.

Our study indicated that ADAM28 was involved in the crown and root morphogen-esis process, and it may participate in this network regulation to link odontogenic cellproliferation and differentiation with matrix synthesis (7). Nevertheless, little is knownabout the relations between ADAM28 and HDPSCs. The objective of this work was toinvestigate the ADAM28 expression features in HDPSCs and the effects of ADAM28gene on the proliferation, differentiation, and apoptosis of HDPSCs via cellular andmolecular biology techniques.

People’s Liberation Army, Beijing, China; and †Department of Stomatology, The 401st Hospital of

ion of China located at Beijing (Project No. 30772450).atology, General Hospital of Chinese People’s Liberation Army, 28 Fuxing Road, Beijing, PR China,

f the American Association of Endodontists.

JOE — Volume 37, Number 3, March 2011

Basic Research—Biology

Materials and MethodsAll procedures were approved by the Ethics Committee on Human

and Animal Research, General Hospital of Chinese People’s LiberationArmy, Beijing, China. The study was conducted from January to July2010. The nucleotide sequence data about human ADAM28 geneappear in the GenBank Nucleotide Sequence Databases under accessionnumber NM-014265.

Subjects and Cell CultureHuman dental pulp has been gently extracted from 40 permanent

third molars of healthy subjects (age range, 18-35 years; 14 women and26 men) at the Department of Stomatology, General Hospital of ChinesePeople’s Liberation Army, from January to July 2010 after writteninformed consent was obtained.

The pulp tissue was separated from the crown and root and thendigested in a solution of 3 mg/mL collagenase type I (Sigma, St Louis,MO) and 4 mg/mL dispase (Gibco-BRL, Grand Island, NY) for 1 hour at37�C as previously described (1). Single-cell suspensions wereobtained by passing the cells through a 70-mm strainer (Falcon; BDLabware, Franklin Lakes, NJ), washed with Dulbecco modified Eaglemedium (DMEM, Gibco-BRL) containing penicillin G (100 U/mL,Gibco-BRL) and streptomycin (100 mg/mL, Gibco-BRL) supple-mented with 15% (v/v) fetal bovine serum (FBS) (Gibco-BRL, GrandIsland, NY), and then cultured in six-well plates (Costar, Cambridge,MA) at 37�C in a humidified atmosphere of 5% CO2 as previouslydescribed (1).

The cells were subcultured by using 0.25% (weight/volume [w/v])trypsin and 0.05% (w/v) EDTA (Invitrogen, Carlsbad, CA) after reaching95% confluency. The second-passage cells were used for the followingisolation.

Cell Isolation and IdentificationFor isolation HDPSCs, an immunomagnetic bead selection method

using the stromatin-1 (STRO-1) antibody was performed as describedpreviously (9). In brief, approximately 2 � 107 cell/mL single-cellsuspensions were incubated with mouse antihuman monoclonal anti-body (mAb) of stromatin-1 (STRO-1, immunoglobulin M; 1:200 dilu-tion; R&D, Chicago, IL) at 4�C for 30 minutes. The cells were washedwith 0.01 mol/L phosphate buffered saline (PBS)/0.1% FBS and incu-bated with rat antimouse immunoglobulin M–conjugated magneticbead (four beads per target cell; Dynal Biotech ASA, Oslo, Norway)for 1 hour on a rotary mixer at 4�C. Sequentially, the bead-bound cellswere isolated by a magnet (Dynal). After washing, the bead-bound cellswere selected using a magnetic particle concentrator (MPC, Dynal) ac-cording to the manufacturer’s recommended protocol. STRO-1–posi-tive cells were counted and harvested for further determination.

The immunofluorescence detections of separated HDPSCs withmAbs against stromatin-1 (STRO-1) and cytokeratin (all at 1:100 dilu-tions, R&D, Chicago, IL) were performed, respectively. The HDPSCsgrowth curve was determined by MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-2H-tetrazoliumbromide) method as described previously(10). Single-cell suspensions of the second-passage HDPCs and purifiedHDPSCs were seeded into a 96-bore plate with 2 � 103 cell/bore, andeach group contains 10 bores. During postinoculation 9 days, opticaldensity (OD) value was determined everyday. The average OD valuewas obtained and used for drawing growth curve.

Subsequently, the cell colony doubling time was detected. TheHDPCs and HDPSCs of log phase growth were digested, counted, andseeded into a 25-cm2 culture flask with 2 � 105 cell/flask at 37�C ina humidified atmosphere of 5% CO2 as previously described (1).Cell counting was performed when the HDPCs and HDPSCs reached

JOE — Volume 37, Number 3, March 2011

log phase again (colony doubling time = t lg2/lg [Nt/N0], t = intervalbetween seeding and assay; N0 = cell population of seeding, Nt = cellpopulation of assay).

Construction of ADAM28 Eukaryotic Expression PlasmidIn this study, human ADAM28 (NM-014265) was cloned from

lymphocytes. A polymerase chain reaction (PCR) product (2,327 bp)corresponding to the total length of ADAM28 coding region (48-2375) was generated using the following gene-specific primers: forward:5’-ATGTTGCAAGGTCTCCTGCCAGTCAGTCTC-3’ (amino acids 48-77) andreverse: 5’-TCATGCTTTTGGATTTGAGTCCTTAGGTGTAGACA-3’ (aminoacids 2375-2341). The PCR product was separated by 1.2% agarosegel electrophoresis, visualized with ethidium bromide staining, and puri-fied by NucleoTrap Gel Extraction Kit (Clontech, Mountain View, CA).

Eukaryotic expression plasmid pcDNA3.1(+)-ADAM28 was con-structed and verified by PCR identification, restriction endonucleasedigestion identification, and DNA sequencing as described previously(10). Meanwhile, glyceraldehyde-3-phosphate dehydrogenase(GAPDH, 900 bp) was used as a control; specific primers were asfollows: forward: 5’-AGCCGCATCTTCTTTTGCGTC-3’ and reverse:5’-TCATATTTGGCAGGTTTTTCT-3’.

Transfection of ADAM28 Eukaryotic Plasmidinto HDPSCs and Expression Detections

Experimental HDPSCs were divided into three groups according toeukaryotic plasmid group, pcDNA3.1(+) group with 30-mmol/Lconcentration, and untransfected group. Transient transfection wasperformed using the Lipofectamine 2000 system (Invitrogen, Carlsbad,CA) as described previously (10). After being transfected for 48 hours,cells on coverslips were washed with PBS, fixed in 4% paraformalde-hyde for 2 hours, and then subjected to immunofluorescence stainingas described previously (11). The other cells were collected and usedfor RT-PCR, western blotting (western blotting kit; Chemicon Interna-tional, Temecula, CA), and cell-cycle detections according to the manu-facturer’s protocol. The polyclonal antibody against ADAM28 (SantaCruz Biotechnology, Santa Cruz, CA) was at 1:100 dilution. GAPDHwas used as an endogenous control. Meanwhile, Labworks Softwareultraviolet photometry gel image analysis system (Cell Biosciences, To-kyo, Japan) was used to detect the grayscales of all PCR bands. AlphaImager 1220 image analysis system (Cell Biosciences, Tokyo, Japan)was used to detect the grayscales of Western blot bands. Relative gray-scale analysis of gene expression was calculated by the DCT method withADAM28/GAPDH, which was normalized to the GAPDH controls.

Design and Transfection of ADAM28Antisense Oligodeoxynucleotides andSense Oligodeoxynucleotides

The nucleotides of 20 nt specifically targeting human ADAM28messenger RNA (GenBank no. NM-014265) were designed and synthe-sized as ADAM28 antisense oligodeoxynucleotide (AS-ODN) (5’-GG CAGGAG ACC TTG CAA CAT-3’, 67-48), meanwhile the sequences of 20 ntwere designed as ADAM28 sense oligodeoxynucleotides (S-ODN) (5’-CAG TCT CCT CCT CTC TGT TG-3’, 71-90). The sequences were subjectedto sulfur modification and fluorescein isothiocyanate (FITC) fluores-cence labeling (Takara, Japan) and confirmed to have satisfactory spec-ificity by National Center for Biotechnology Information/Basic LocalAlignment Search Tool 2.0 database search. The experimental concen-tration was 30 mmol/L. Transfection procedures were the same asmentioned previously. Experimental cells were divided into three groupsaccording to the ADAM28 AS-ODN group, the S-ODN group with 30mmol/L transfection concentration, and the untransfected group.

The Biologic Effects of ADAM28 on HDPSCs 333

Basic Research—Biology

Transfection efficiency was observed under fluorescence microscopeafter being transfected for 48 hours, and the transfection efficiency(%) = fluorocyte number at the same eyeshot/total cellular score atthe same eyeshot under inverted phase contrast microscope �100%.The inhibitory effect of ADAM28 AS-ODN on HDPSCs was determinedby RT-PCR and western blot assays. The relative grayscale values fromRT-PCR and western blot assays of each group were obtained, and theresults were subjected to statistical analysis. To detect the effects ofADAM28 on biological characteristics of HDPSCs after being transfectedfor 48 and 72 hours, cell biology and enzyme kinetics assays were per-formed, and the HDPSC density was retained at 1 � 106/mL.

Cell Proliferation Assay by MTTThe cell proliferation assay was performed in all five groups of

HDPSCs using the MTT method as described previously (10). Briefly,20 mL MTT (5 g/L; Sigma-Aldrich, St. Louis, MO) was added to eachwell and cells were incubated for 4 hours at 37�C. Supernatant ineach well was then removed, and 150 mL dimethyl sulphoxide (Sigma)was added for 10 minutes. Cell proliferation was detected quantitativelyusing a microplate enzyme-linked immunosorbent assay reader with anOD value of 490 nm. The assay was repeated for five times, and the datawere presented as the mean � standard deviation.

Cell-cycle Detection by Flow CytometryTo detect the proliferative activity of HDPSCs, the cellular DNA

content from each stage of five groups was determined by flow cytometryas described previously (12). We collected the attached cells usingtrypsin-EDTA and resuspended them in DMEM (no FBS). The cellswere fixed for 30minutes in an ice-cold 70% ethanol solution containingribonuclease (RNase, 2 mg/mL). We washed the cells in PBS and thenstained themwith propidium iodide (PI) for 30minutes at room temper-ature in the dark. The PI-elicited fluorescence was measured for indi-vidual cells using a FACSCalibur flow cytometer (Becton Dickinson,Tokyo, Japan) with laser excitation at 488 nm. We analyzed a total of2 � 106 cells for each sample and determined the percentages of cellsin G0/G1, S, and G2+M phases using standardModiFit and Cell Quest soft-ware (Becton Dickinson Biosciences, San Diego, CA).

Alkaline Phosphatase Activity DetectionAfter 48 and 72 hours of transfection, respectively, the HDPSCs of

all five groups were washed with 0.01 mol/L PBS and 50 mL of cold 10mmol/L tris-HCl buffer (pH = 7.4) containing 0.1% TritonX-100 wasadded before incubation at 4�C overnight. Alkaline phosphatase(ALP) substrate solution (100 mL) containing 2 mmol/L MgCl2 and16 mmol/L p-nitrophenyl phosphate was then mixed with each sample.After incubation for 30 minutes at 37�C, the reaction was stopped byadding 50 mL of 0.2 mol/L NaOH, and the liberated p-nitrophenolwas measured spectrophotometrically at 410 nm. The change in rateof absorbance was directly proportional to ALP activity. Each experi-ment was repeated five times, and the data were presented as themean � standard deviation.

Dentin Sialophosphoprotein, Dentin Matrix Protein 1, andBone Sialoprotein

Immunocytochemical (LsABC kit; DAKO Ltd, Cambridgeshire, MA)staining, western blotting, and image analysis were performed betweenthe eukaryotic plasmid group, the pcDNA3.1(+) group, and the untrans-fected group. Polyclonal antibody against dentin sialophosphoprotein(DSPP), dentin matrix protein 1 (DMP1), and bone sialoprotein(BSP) (Santa Cruz Biotechnology, Santa Cruz, CA) were at 1:200 dilution.Control experiments were performed by replacing the primary antibody

334 Zhao et al.

with nonimmune rabbit serum. The average grayscales (mean � stan-dard deviation) of each group were subjected to statistical analysis.

Apoptosis Analysis by Annexin V-FITC/PI AssayThe HDPSCs of the eukaryotic plasmid group, the pcDNA3.1(+)

group, and the untransfected group were cultured in DMEM (no FBS)for 4 hours before the transfection procedure. After 48 and 72 hoursof transfection, the HDPSCs were harvested, washed with cold PBS,and stained with PI and FITC-conjugated annexin V using an AnnexinV-FITC Apoptosis Detection Kit I (Becton Dickinson). Annexin V-FITCidentifies cells in early apoptosis by detecting externalized phosphatidyl-serine, and PI identifies cells that have lost plasma membrane integrity(ie, necrotic or late apoptotic cells). The HDPSCs were resuspendedin 50 mL of 1� binding buffer supplemented with 5 mL of annexinV-FITC and 10 mL of PI and kept at room temperature in the dark for15 minutes according to the manufacturer’s instructions. After the addi-tion of 450 mL of 1� binding buffer, the stained cells were kept on iceand subjected to fluorescence-activated cell sorter (FACS) analysis usinga FACSCalibur flow cytometer (BD Biosciences, San Diego, CA) with CellQuest software (BD Biosciences). The FITC fluorescence was between515 and 545 nm, and the PI fluorescence was between 564 and 606 nm.

Statistical AnalysisResults were analyzed and expressed as themean� standard devi-

ation. The statistical differences between various groups were evaluatedby the Student-Newman-Keuls test from SPSS Windows version 12.0program (SPSS Inc, Chicago, IL). Comparisons were considered signif-icant if p < 0.05.

ResultsThe Morphology and Source Identification of HDPSCs

The STRO-1+ HDPSCs, which represented approximately 6% of thetotal pulp cell population, were separated. Purified HDPSCs displayeda short fusiform shape, and its nucleus was located at the center ofthe cytoplasm and presented an orbicular-ovate, vacuoles shape; thenucleoli were clear (Fig. 1A). The shapes and growth property ofHDPSCs from four groups after transfection had no significant differ-ence compared with those without transfection, the minority of HDPSCsappeared suspended or dead, and the attached total cellular score wasmaintained at 1 � 106/mL.

Immunofluorescence detection indicated that STRO-1 (Fig. 1B)was strongly positive in cytoplasms and cytomembranes of HDPSCs,whereas cytokeratin (Fig. 1C) was negative in HDPSCs, which showedthat all isolated HDPSCs were derived from the mesenchyme with nocontamination from odontogenic epithelial cells. Cell growth curve(Fig. 1D) displayed that the second-passage HDPSCs and HDPCs residedin transient growth retardation phase between days 1 and 3. On day 4,the cell proliferation speed accelerated and got into log phase, and up today 8, the proliferation of the HDPSCs and HDPCs peaked. The cellcolony doubling time of the HDPSCs and HDPCs during days 4 and 8was 3.92 and 3.68 days, respectively. The results indicated that theproliferation velocity of purified HDPSCs by STRO-1 was slower thanthat of unpurified HDPCs, which is in accordance with stem cell featureof slowly proliferative cycle.

The Expression Detections after Transfection of ADAM28Eukaryotic Plasmid

The immunofluorescence assay indicated that robust green fluo-rescence appeared in pcDNA3.1(+)-ADAM28 group, whereas no fluo-rescence was seen in the other two groups (Fig. 2A). RT-PCR detection

JOE — Volume 37, Number 3, March 2011

Figure 1. The morphology and source identification of HDPSCs. (A) The purified HDPSCs with a short fusiform shape. (B) STRO-1 was strongly positive in cyto-plasms and cytomembranes of HDPSCs by FITC fluorescence. (C) Cytokeratin was negative in cytoplasms and cytomembranes of HDPSCs, whereas diamidino-phenyl-indole was expressed in nucleus of HDPSCs. Scale bar = 20 mm. (D) The cell growth curve indicated that the proliferation cycle of purified HDPSCswas longer than that of unpurified HDPCs, which is in accordance with stem cell feature.

Basic Research—Biology

showed that the corresponding brightness of the GAPDH (900 bp) bandfor each group was almost the same, whereas a specific bright band of2,327 bp representative of ADAM28 messenger RNA was solely ex-pressed in the eukaryotic plasmid group (Fig. 2B).

Western blotting detection revealed that the GAPDH (36 kD) bandfor each group was almost the same, but a clear protein band of 35.3 kD(ADAM28) was only found in the eukaryotic plasmid group (Fig. 2C).These results confirmed the high efficiency of the present transfectionsystem and ensured the role of ADAM28 manipulation for further anal-ysis of biological property of HDPSCs.

Transfection Efficiency and Biological Effect of ADAM28AS-ODNs on HDPSCs

Significant immunofluorescence was observed in cytoplasms andnucleoli of the AS-ODN and S-ODN groups under a microscope. Thetransfection efficiency of the AS-ODN and S-ODN groups was 85%and 81%, respectively, and there was no significant difference betweenthe two groups (Fig. 3A).

Figure 2. The expression detections after the transfection of ADAM28 eukaryoticplasmid group via immunofluorescence assay, whereas no fluorescence was seenthat a specific bright band of 2,327 bp indicative of ADAM28 messenger RNA wasrevealed that a clear protein band of 35.3 kD (ADAM28) was only seen in eukary

JOE — Volume 37, Number 3, March 2011

An inhibitory effect was detected by RT-PCR and western blotting.The RT-PCR result indicated that the ADAM28 (255 bp) band of the AS-ODN group was brighter than that of the S-ODN group. Statistical anal-ysis showed that the ADAM28 expression level of the AS-ODN group washigher than that of the S-ODN group, and a significant difference wasdetected (p < 0.01) (Fig. 3B). Western blotting analysis revealed thata 35.3-kD protein band in the AS-ODN group was notably strongerthan that of the S-ODN group, and there was a significant differencebetween the two groups (p < 0.01) (Fig. 3C). These results suggestedthat the ADAM28 AS-ODN had a positive enhancing effect on ADAM28expressions in HDPSCs.

Influence of ADAM28 Eukaryotic Plasmid and AS-ODNs onBiological Characteristics of HDPSCs

The MTT assay showed that the ADAM28 AS-ODN group exhibitedthe highest OD value at 48 and 72 hours, whereas in the ADAM28 eu-karyotic plasmid group cell proliferation was markedly inhibited

plasmid. (A) ADAM28 messenger RNA was strongly expressed in eukaryoticin the other two groups. Scale bar = 20 mm. (B) RT-PCR detection displayedonly expressed in eukaryotic plasmid group. (C) Western blotting detectionotic plasmid group.

The Biologic Effects of ADAM28 on HDPSCs 335

Figure 3. The transfection efficiency and inhibitory effect of ADAM28 AS-ODN on HDPSCs. (A) Strongly immunofluorescence was found in cytoplasms and nucleoliof the AS-ODN and S-ODN groups at the almost same level. Scale bar = 20 mm. (B) RT-PCR detection indicated that ADAM28 (255 bp) band of the AS-ODN groupwas brighter than that of the S-ODN group, whereas no expression was found in untransfected group. Statistical analysis showed that ADAM28 expression level of theAS-ODN group was much higher than that of the S-ODN group, and a significant difference was detected (p < 0.01). (C) Western blotting showed that a 35.3-kDprotein band in the AS-ODN group was notably stronger than that of the S-ODN group, and there was a significant difference between the two groups (p < 0.01).Quantitative data are expressed as the means � standard deviation.

Basic Research—Biology

(Fig. 4A). The results indicated that ADAM28 had a negative regulatoryeffect on the HDPSCs proliferation.

Cell-cycle distributions revealed that the ADAM28 AS-ODN groupshowed the highest percentage of cells in S/G2+M phases, whereastransfection of ADAM28 eukaryotic plasmid resulted in the lowestpercentage of cells in S/G2+M phases compared with the other groups(Fig. 4B, p < 0.01), suggesting that ADAM28 eukaryotic plasmid couldsignificantly inhibit the HDPSCs proliferation.

In contrast to the cell proliferation assay, ALP activity analysis dis-played a reverse trend that ADAM28 AS-ODN group produced the lowestvalue of ALP absorbance, whereas eukaryotic plasmid-transfectedHDPSCs kept the highest level compared with the untransfected groupafter 48 or 72 hours of culture (Fig. 4C). The results indicated thatADAM28 AS-ODN could significantly inhibit ALP activity of HDPSCs.

Immunocytochemical staining indicated that DSPP was stronglypositive in cytoplasms of the eukaryotic plasmid group, whereasa weaker positive was shown in the pcDNA3.1(+) group and the un-transfected group (Fig. 4D). Image analysis indicated that there wasa significant difference between the eukaryotic plasmid group and theother groups (Fig. 4E, *p < 0.05). Western blotting and quantitativeanalysis revealed that the DSPP expression level (110 kD) of the eukary-otic plasmid group was markedly higher than that of the other twogroups, and a significant difference was found between the eukaryoticplasmid group and the other groups (Fig. 4F, p < 0.01). These resultsshowed that the transfection of ADAM28 eukaryotic plasmid couldsignificantly promote the DSPP expression in HDPSCs, and ADAM28

336 Zhao et al.

has a positive correlation with DSPP although no direct relationswere found between ADAM28 and DMP1/BSP (data not shown).

Cell-apoptosis assay was performed to detect whether this effect wasrelated to the changes in cell survival; the results suggested that therewere many more apoptotic cells and fewer live cells in the ADAM28 eu-karyotic plasmid group than in the other groups, and there was a signif-icant difference between any two groups (Fig. 4G and H, p < 0.01).

DiscussionDPSCs were described as mesenchymal stem cell–like

odontogenic precursor cells with highly proliferative potential able toregenerate dentin in an immunocompromised host (1). Previousstudies indicated that when HDPSCs were transplanted into immuno-compromised mice, they generated a dentine-like structure lined withodontoblast-like cells that surrounded a pulp-like interstitial tissue(13). Further analyses showed that DPSCs were able to differentiateinto specialized cell types other than odontoblasts (3). The selectionof dental pulp stem cells from the third molar can be an easy accessiblechannel for this study. Many wisdom teeth cannot erupt at the appro-priate position and stay impacted because of inadequate jawbone space(14).

In this study, 6% STRO-1–positive HDPSCs were successfully sepa-rated using immunomagnetic bead selection, which has been a moreeffective method for the isolation, purification, and identification ofmesenchymal stem cell populations (9). Using STRO-1–bound

JOE — Volume 37, Number 3, March 2011

Figure 4. The influence of diverse ADAM28 expression on proliferation, differentiation, and apoptosis of HDPSCs. (A) The MTT assay displayed that the ADAM28AS-ODN group exhibited the highest OD value at 48 and 72 hours, whereas in the ADAM28 eukaryotic plasmid group cell proliferation was markedly inhibited ata statistically significant level. Data are presented as the means � standard deviation from five groups.*p < 0.01 versus other groups. (B) Cell-cycle distributionsdisclosed that the ADAM28 AS-ODN group showed the highest percentage of cells in S/G2+M phases (p < 0.01), whereas the transfection of ADAM28 eukaryoticplasmid resulted in the lowest percentage of cells in S/G2+M phases compared with the other groups (p < 0.01). (C) ALP activity analysis indicated a reverse trendthat the ADAM28 AS-ODN group produced the lowest value of ALP absorbance, whereas eukaryotic plasmid-transfected HDPSCs kept the highest level after 48 or 72hours of culture. Data are shown as the means� standard deviation from five groups. *p < 0.01 versus other groups. (D) Immunocytochemistry showed that DSPPwas strongly positive in cytoplasms of the eukaryotic plasmid group, whereas a weaker positive was shown in the pcDNA3.1(+) group and untransfected groups. (E)Statistical analysis indicated that a significant difference was found between the eukaryotic plasmid group and the other groups (p < 0.05). (F) Western blotting andquantitative analysis showed that the DSPP expression level of the eukaryotic plasmid group was markedly higher than that of other two groups, and there wasa significant difference between eukaryotic plasmid group and the other groups (*p < 0.01). (G and H) The cell-apoptosis result showed that there weremore apoptotic cells and fewer live cells in the ADAM28 eukaryotic plasmid group than in the other groups, and a significant difference was detected betweenany two groups (*p < 0.01).

Basic Research—Biology

JOE — Volume 37, Number 3, March 2011 The Biologic Effects of ADAM28 on HDPSCs 337

Basic Research—Biology

magnetic beads, HDPSCs can be efficiently retrieved from single cellsuspensions. This is based primarily on the notion that DPSCs may orig-inate from perivascular cells migrating out of capillary walls into thesurrounding fibrous pulp tissue in response to the degradation ofdentin matrix (15). The isolation of clonogenic and multilineage poten-tial cells from diverse tissues has qualified STRO-1 antibody as a suitablemarker for different mesenchymal stem cell populations (16).

It has been shown that ADAM28, as a possible virulence gene forCHTR, is involved in various biological events such as cell adhesion, cellfusion, cell migration, membrane protein shedding, and proteolysis(4). According to the clinical situation, congenital hypoplasia of toothroot (CHTR) is separated into cementum developmental defect of toothroot, dentin developmental anomaly of tooth root, root paramorphia(absence, short cone shape), and root adhesive organ dysplasia.CHTR could be occurred in deciduous and permanent dentistry,whereas its crown growth and dental eruption are normal (8). Accord-ingly, CHTR has been proposed to correlate with the proliferation anddifferentiation disorder of odontogenic mesenchymal cells.

In the study, the effects of ADAM28 on the biologic functions ofHDPSCs were determined. Firstly, ADAM28 eukaryotic plasmid andAS-ODNs were constructed and transfected into HDPSCs successfully.The detections indicated that ADAM28 messenger RNA and proteinwere solely transcribed, translated, and expressed in the eukaryoticplasmid group, and AS-ODN promoted the ADAM28 expression inHDPSCs, which was opposite to inhibitory potency of AS-ODN. There-fore, we have proposed that AS-ODN might have a positive-feedbackregulatory role on ADAM28 distribution in HDPSCs by coordinationwith other signaling molecules and lead to a stronger expression ofADAM28.

We have also shown that ADAM28 AS-ODN could significantlyenhance the HDPSCs proliferation, whereas the eukaryotic plasmidgroup could exert the opposite effect, which suggested that ADAM28might play a negative regulatory role in the HDPSCs proliferation byMTT assay and cell-cycle detection. Cell-cycle detection displayed thatthe maximal cells accumulated in the S phase (DNA synthesis phase)of the AS-ODN group, which resulted in the elevation of the total PI value(cell proliferation index, S+G2+M); meanwhile, the eukaryotic plasmidgroup exhibited the decrease of PI value compared with those of thecontrol groups. The critical step required for the proliferation of cellsis their transition from cell-cycle arrest to entry into the active phase ofthe cell cycle through the G1/S restriction point; thus, the fast transitioncan elevate higher rates of cell proliferation (16). These descriptionswould prove that ADAM28 could depress the HDPSCs proliferation bystepping down the transition process from G1 to S phase, further inhib-iting DNA reproduction and protein synthesis of HDPSCs.

We analyzed the action mechanism could be that ADAM28comprises an EGF-like and metalloproteinase-like domain, whichmay possess catalytic activity and certain functions of EGF such as themaintenance of undifferentiated cell proliferation (17). Cell apoptosisand proliferation are related and interact with each other in tooth devel-opment, and an understanding of apoptosis regulation in the vestigialtooth primordia can help to elucidate the mechanism of their suppres-sion during evolution (18). Thus, apoptosis may be a general mecha-nism for the silencing of embryonic signaling centers includingodontogenesis signal transmission (19). Our results revealed thatmore apoptotic cells and fewer live cells were found in the ADAM28 eu-karyotic plasmid group than in the other groups. ADAM28 eukaryoticplasmid could notably not only promote apoptosis of HDPSCs butalso change the characteristics of live cells in terms of cell apoptosis,which strongly suggested that ADAM28 might actively participate inthe positive-feedback mechanism of the apoptotic event during toothmorphogenesis.

338 Zhao et al.

In the developing tooth, the presence of ALP is concomitant withthe onset of cell differentiation (20) and is believed to be prerequisitefor pulp cells to differentiate into odontoblasts (21). ALP activity isuniversally regarded as a crucial enzymemarker of odontogenic mesen-chymal cytodifferentiation andmatrix mineralization (22). In this study,the overexpression of ADAM28 enhanced the HDPSCs differentiation,whereas the use of ADAM28 AS-ODN produced the opposite effect, asshown by ALP absorbance.

The previously described evidence indicated that ADAM28 mightactively participate in the proliferation, apoptosis, and differentiationstart of HDPSCs. Cytodifferentiation expressed by ALP activity and cellproliferation evaluated by the MTT and cell-cycle analysis coincidentallyshowed that ADAM28 could inhibit HDPSCs proliferation during theprocess of differentiation enhancement, which was in accordance withthe previous concept that proliferation and differentiation are oppositelycorrelated with each other because of the existence of dual-functionregulators involved in controlling both the processes (23). Therefore,it is reasonably proposed that some components of ADAM28 couldact not only as inhibitors for the HDPSCs proliferation but also aspromoters for cytodifferentiation.

Together with ALP, DSPP, DMP1, BSP, and osteopontin have longserved as unique markers of odontoblasts differentiation during odon-togenesis and believed to play an essential role in dentin formation andremodeling (24). Dentin sialoprotein and dentin phosphoprotein areconsidered as specific markers for terminally differentiated odonto-blasts and shown to be cleavage products of a primary transcriptprecursor encoded by a single gene termed DSPP on human chromo-some 4 (25). DSPP was first cloned from developing teeth and thoughtto be tooth specific (26). Moreover, DSPP may be implicated in theinitial mineralization process of the dentin matrix collagen (27) basedon the defective dentin formation in DSPP-null mice (28). The expres-sion levels of DSPP were directly correlated with a dentin developmentaldefect of tooth root. In this study, ADAM28 played a positive regulatoryrole in the expression level of DSPP in HDPSCs, which suggested that theADAM28 gene could significantly promote the functional differentiationof HDPSCs into odontoblasts and further enhance dentin development.

During tooth development, as an acidic phosphorylated extracel-lular matrix protein, DMP1 is expressed in the dental pulp cells, odon-toblasts, predentin, dentin, and cementum. Moreover, DMP1 is essentialin the maturation of odontoblasts and osteoblasts as well as in miner-alization via local and systemic mechanisms (29). BSP is a major miner-alized tissue-specific noncollagenous extracellular matrix protein that ishighly glycosylated, phosphorylated, and sulfated (30). The temporo-spatial deposition of BSP into the extracellular matrix of bone andthe ability of BSP to nucleate hydroxyapatite crystal formation (31) indi-cate a potential role for BSP in the initial mineralization of bone, dentin,and cementum. Therefore, regulation of the BSP gene is potentiallyimportant in the differentiation of osteoblasts, odontoblasts, cemento-blasts, and ameloblasts (32) and in bone matrix mineralization.

In the study, the transfection of ADAM28 had no striking effect onthe expression levels of DMP1 and BSP, which revealed that ADAM28might not notably facilitate the differentiation and maturation of osteo-blasts and cementoblasts, and furthermatrix mineralization of bone andcementum, namely the ADAM28 gene, had no intimate correlations withthe biologic property of DMP1 and BSP in HDPSCs.

Compared with our previous researches, the biological function ofthe HDPSCs in this study was almost contrary to that of the human peri-odontal ligament stem cells (HPDLSCs) (33). The HPDLSCs studies indi-cated that after transfected into HPDLSCs, the ADAM28 eukaryoticplasmid group showed the highest expression level, whereas the AS-ODN group displayed the lowest. Furthermore, overexpression ofADAM28 enhanced the HPDLSCs proliferation and inhibited specific

JOE — Volume 37, Number 3, March 2011

Basic Research—Biology

differentiation of HPDLSCs, whereas the inhibition of ADAM28 producedthe opposite effects and induced apoptosis, which suggested thatADAM28 might have a positive regulatory effect on the proliferation ofHDPSCs and present negative correlations with differentiation andapoptosis of HPDLSCs. Moreover, ADAM28 AS-ODN could inhibitcementum attachment protein expression, and ADAM28 has a positivecorrelation with the cementum attachment protein (33). From the studymentioned above (33), we analyzed that the mechanismmight be that theHPDLSCs and HDPSCs originally come from the dental follicle cells anddental papilla cells, respectively, whose biologic characteristics anddifferentiation trends have striking differences. Meanwhile, theADAM28 expression distribution in diverse odontogenic cells was alsodifferent, which could lead to the developmental defect of tooth root.

The present information enlightened us that we might attempt totransfect ADAM28 eukaryotic plasmid into the HDPSCs to accelerateits functional differentiation into odontoblasts and further producedentin to cure dentin developmental defect of tooth root and root para-morphia resulted from CHTR. In conclusion, this study testifies convinc-ingly the expression distribution of ADAM28 in HDPSCs and confirmsthat ADAM28 could effectively manipulate the proliferation, differentia-tion, and apoptosis of HDPSCs at transcription, translation, and proteinlevels. In the future, we would use distinct expression characteristics ofADAM28 for gene therapy against dentin developmental defect of toothroot and root paramorphia, which will provide a new pathway for clin-ical treatment against CHTR.

AcknowledgmentsThe authors deny any conflicts of interest related to this study.

References1. Gronthos S, Mankani M, Brahim J, et al. Postnatal human dental pulp stem cells

(DPSCs) in vitro and in vivo. Proc Natl Acad Sci U S A 2000;97:13625–30.2. Stokowski A, Shi S, Sun T, et al. EphB/ephrin-B interaction mediates adult stem cell

attachment, spreading, and migration: implications for dental tissue repair. StemCells 2007;25:156–64.

3. About I, Bottero MJ, de Denato P, et al. Human dentin production in vitro. Exp CellRes 2000;258:33–41.

4. Ohtsuka T, Shiomi T, Shimoda M, et al. ADAM28 is overexpressed in human non-small cell lung carcinomas and correlates with cell proliferation and lymph nodemetastasis. Int J Cancer 2006;118:263–73.

5. McLane MA, Marcinkiewicz C, Vijay-Kumar S, et al. Viper venom disintegrins andrelated molecules. Proc Soc Exp Biol Med 1998;219:109–19.

6. Mitsui Y, Mochizuki S, Kodama T, et al. ADAM28 is overexpressed in human breastcarcinomas: implications for carcinoma cell proliferation through cleavage ofinsulin-like growth factor binding protein-3. Cancer Res 2006;66:9913–20.

7. Zhao Z, Wen LY, Jin M, et al. ADAM28 participates in the regulation of tooth devel-opment. Arch Oral Biol 2006;51:996–1005.

8. Apajalahti S, Arte S, Pirinen S. Short root anomaly in families and its association withother dental anomalies. Eur J Oral Sci 1999;107:97–101.

9. Yu J, Deng Z, Shi J, et al. Differentiation of dental pulp stem cells into regular-shapeddentin-pulp complex induced by tooth germ cell conditioned medium. Tissue Eng2006;12:3097–105.

JOE — Volume 37, Number 3, March 2011

10. Zhao Z, Tang L, Deng ZH, et al. Essential role of ADAM28 in regulating the prolif-eration and differentiation of human dental papilla mesenchymal cells (hDPMCs).Histochem Cell Biol 2008;130:1015–25.

11. Zhao Z, Liu HC, Jin Y, et al. Influence of ADAM28 on biological characteristics ofhuman dental follicle cells. Arch Oral Biol 2009;54:835–45.

12. Zhang W, Walboomers XF, Shi S, et al. Multilineage differentiation potential of stemcells derived from human dental pulp after cryopreservation. Tissue Eng 2006;12:2813–23.

13. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult humanmesenchymal stem cells. Science 1999;284:143–7.

14. Kruger E, Thomson WM, Konthasinghe P. Third molar outcomes from age 18 to 26:findings from a population-based New Zealand longitudinal study. Oral Surg OralMed Oral Pathol Oral Radiol Endod 2001;92:150–5.

15. Shi S, Gronthos S. Perivascular niche of postnatal mesenchymal stem cells in humanbone marrow and dental pulp. J Bone Miner Res 2003;18:696–704.

16. Santamaria D, Ortega S. Cyclins and CDKs in development and cancer: lessons fromgenetically modified mice. Front Biosci 2006;11:1164–88.

17. Howard L, Zheng Y, Horrocks M, et al. Catalytic activity of ADAM28. FEBS Lett 2001;498:82–6.

18. Peterkova R, Peterka M, Lesot H. The developing murine dentition: a new tool forapoptosis study. Ann N Y Acad Sci 2003;1010:453–66.

19. Vaahtokari A, Aberg T, Thesleff I. Apoptosis in the developing tooth: association withan embryonic signaling center and suppression by EGF and FGF-4. Development1996;122:121–9.

20. Woltgens JH, Lyaruu DM, Bronckers AL, et al. Biomineralization during early stagesof the developing tooth in vitro with special reference to secretory stage of amelo-genesis. Int J Dev Biol 1995;39:203–12.

21. Unda FJ, Martin A, Hernandez C, et al. FGFs-1 and-2, and TGFb1 as inductive signalsmodulating in vitro odontoblast differentiation. Adv Dent Res 2001;15:34–7.

22. Shiba H, Mouri Y, Komatsuzawa H, et al. Enhancement of alkaline phosphatasesynthesis in pulp cells co-cultured with epithelial cells derived from lower rabbitincisors. Cell Biol Int 2003;27:815–23.

23. Zhu L, Skoultchi AI. Coordinating cell proliferation and differentiation. Curr OpinGenet Dev 2001;11:91–7.

24. Huang W, Yang S, Shao J, et al. Signaling and transcriptional regulation in osteoblastcommitment and differentiation. Front Biosci 2007;12:3068–92.

25. MacDougall M, Simmons D, Luan X, et al. Dentin phosphoprotein and dentin sialo-protein are cleavage products expressed from a single transcript coded by a gene onhuman chromosome 4. Dentin phosphoprotein DNA sequence determination. J BiolChem 1997;272:835–42.

26. D’Souza RN, Cavender A, Sunavala G, et al. Gene expression patterns of murinedentin matrix protein 1 (Dmp1) and dentin sialophosphoprotein (DSPP) suggestdistinct developmental functions in vivo. J Bone Miner Res 1997;12:2040–9.

27. Butler WT. Dentin matrix proteins. Eur J Oral Sci 1998;106:204–10.28. Sreenath T, Thyagarajan T, Hall B, et al. Dentin sialophosphoprotein knockout

mouse teeth display widened predentin zone and develop defective dentin mineral-ization similar to human dentinogenesis imperfecta type III. J Biol Chem 2003;278:24874–80.

29. Qin C, D’Souza R, Feng JQ. Dentin matrix protein 1 (DMP1): new and importantroles for biomineralization and phosphate homeostasis. J Dent Res 2007;86:1134–41.

30. Ogata Y, Yamauchi M, Kim RH, et al. Glucocorticoid regulation of bone sialoprotein(BSP) gene expression. Identification of a glucocorticoid response element in thebone sialoprotein gene promoter. Eur J Biochem 1995;230:183–92.

31. Hunter GK, Goldberg HA. Nucleation of hydroxyapatite by bone sialoprotein. ProcNatl Acad Sci U S A 1993;90:8562–5.

32. Paz J, Wade K, Kiyoshima T, et al. Tissue and bone cell-specific expression of bonesialoprotein is directed by a 9.0 kb promoter in transgenic mice. Matrix Biol 2005;24:341–52.

33. Zhao Z, Wang Y, Wang D, et al. The regulatory role of a disintegrin and metallopro-teinase 28 on the biologic property of human periodontal ligament stem cells.J Periodontol 2010;81:934–44.

The Biologic Effects of ADAM28 on HDPSCs 339